Hello Giovanni,
Great to hear that everything is working much better for you now and that
everything is much faster and smaller than NPY ;)
Do you know how the default value is set btw?
This is computed via a magical heuristic algorithm written by Francesc (?)
called computechunksize().
This i
Hi Anthony,
thanks for the explanation and the links, it's much clearer now. So without
compression a CArray is really a smarter type of sparse file, but you have to
set a sensible chunk shape. Do you know how the default value is set btw? I am
asking because I didn't see any change in performa
Hi Giovanni!
I think that you may have some misunderstanding about how chucking works,
which is leading you to get terrible performance. In fact what you
describe is a great strategy (right all and zip) for using normal Arrays.
However, chunking and CArrays don't work like this. If a chunk cont
Hi all,
I have a sparse 3.4M x 3.4M adjacency matrix with nnz = 23M and wanted
to see if CArray was an appropriate solution for storing it. Right now I
am using the NumPy binary format for storing the data in coordinate
format and loading the matrix with Scipy's sparse coo_matrix class. As
far
